Intel GPU AI Skills is a collection of agent skills for setting up, running, benchmarking, and profiling Hugging Face models on Intel GPUs. It supports workflows involving PyTorch, vLLM-XPU, SGLang-XPU, llama.cpp-SYCL, and migration from CUDA to XPU. The catalogue contains the project's skills, instructions, agent, and plugin.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add intel/gpu-ai-skills --skill model-config-recommendgit clone --depth 1 https://github.com/intel/gpu-ai-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/intel/gpu-ai-skills/model-config-recommend)<a href="https://agentmods.dev/skills/intel/gpu-ai-skills/model-config-recommend"><img src="https://agentmods.dev/badge/skills/intel/gpu-ai-skills/model-config-recommend.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.03646 |
| Opus 5 | $0.00000 | $0.01823 |
| Sonnet 5 | $0.00000 | $0.00729 |
| Haiku 4.5 | $0.00000 | $0.00365 |
Grade A, and why
model-config-recommend scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 275 lines — stays where its author put it; the contents beside it link to each section on GitHub.
model-config-recommend
Status: experimental. Emits physics-bounded predictions, not measurements. Scoped to vLLM-XPU; for SGLang use model-can-it-fit + sglang-xpu-run + sglang-xpu-bench.
When to use
The user has chosen an HF decoder-only LLM and wants to know which vLLM-XPU config (quant, KV dtype, DP/TP, max concurrency, max context) to start with on Intel Arc B-series. Broad wording such as "How should I configure vLLM for this model on my Arc cards?" still activates this skill because it requires selecting layout and deployment parameters rather than merely launching a server. Skip when:
- Exact throughput numbers required -> use vllm-xpu-bench.
- VLM or diffusion model -> use model-can-it-fit.
- Hardware not in
data/hardware.json(other Intel families). - MoE expert parallelism or speculative-decoding speedup — both workload-specific; the skill flags them as bench-only.
Mandatory output contract (read first)
Every recommendation answer MUST do all three, even when the model fits on one GPU — there is no "it's small, skip this" exception:
- Run
recommend.pythis session against the user's actual model, device, context, and concurrency. Never answer from memory or from the "Worked example" numbers below. - State the layout as the literal
dp=N, tp=Mtoken (e.g.dp=1, tp=1). Prose like "no TP needed" does not count. - Reproduce the full
docker run ... <model> ...launch block verbatim in a fenced code block. Never a flag table, a barevllm serveline, or a "block above" pointer instead.
"Reporting the recommendation" below has the detail.
Three tiers
| Tier | Script | What it does |
|---|---|---|
| 1 | recommend.py |
Pulls config.json, applies roofline math against the spec table, emits candidates + launch line. Stdlib only, no GPU. ~2s. |
| 2 | calibrate.py |
Runs a short BF16 bench on a reference model (Qwen/Qwen2.5-1.5B-Instruct by default), measures actual MFU/BWE, caches per (image, device). Requires Docker. |
| 3 | verify.py |
Launches the recommended config on the target model, prints predicted-vs-measured with IN BAND / OUT OF BAND flags. |
What ships with it
7 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 275 lines · 0 tokens per session scan A b112c26118b2
model-config-recommend is a skill published in the GitHub repository intel/gpu-ai-skills (21 stars, last pushed 2d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,646 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
spark-environment-setup
Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). Use when installing PyTorch/Unsloth/TRL/vLLM on DGX Spark, hitting libcudart or wheel-ABI errors on aarch64, or choosing between NGC containers and bare pip installs.
spark-memory-thermal-ops
Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.
spark-training-gotchas
Preflight and diagnose the ten known failure modes for ML training on NVIDIA DGX Spark. Use when a training run on DGX Spark fails to start, OOMs below the 128GB limit, slows down mid-run, or before any multi-hour training job on GB10.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
amc-run-rtsp-calibration
Calibrate a new dataset from live RTSP camera streams via the AutoMagicCalib REST API. Use when the user provides RTSP URLs or asks to calibrate live cameras; VIOS records clips, AMC ingests them, then runs calibration.
amc-run-video-calibration
Calibrates pre-recorded cam.mp4 datasets through the AutoMagicCalib REST API. Use for user-supplied local MP4s; route live RTSP streams to amc-run-rtsp-calibration.